Head Start Moves Ahead--Without National Testing

WaPo article over the weekend says Congress's Head Start reauthorization will likely discontinue the National Reporting System, a Bush administration-created test of all Head Start four- and five-year-olds that has been very controversial. Under the NHS, several small tests, measuring English language skills, vocabulary, letter recognition and early math skills, are administered to all four- and five-year-olds in Head Start every spring and fall, to assess how much Head Start programs are helping children learn. This is not a bad idea: Accountability has become a key education reform principle, policymakers are increasingly focused on early learning and therefore increasingly concerned about the educational aspect of Head Start, and there are concerns about the quality of some Head Start programs. All these factors support an increased effort to collect more systematic information about student learning in Head Start programs.

But the NRS has lots of problems. If you think educational accountability is complicated at the K-12 level, it's much more so with preschool-aged youngsters, whose development is more sporadic and whose skills are less stable. The NRS was implemented in a rush and there were a number of implementation problems. It's comprised of snippets from several other tests that are valid, but the NRS itself has never been validated. It's focused on a narrow range of early literacy and math skills and does not include important issues of social, emotional and motor development. Some questions have problems of class/cultural/racial bias, are developmentally inappropriate, or measure the wrong things. It doesn't include accommodations for students with disabilities, even though Head Start programs are mandated to serve 10% students with disabilities, and there are also problems with its treatment of non-English speaking children.

On top of these practical and technical concerns, which have been raised by such uninterested parties as the GAO, you have predictable political opposition to the NRS from people who oppose testing and academic instruction for young children on ideological grounds, from Head Start interests that do not want to be held accountable for their performance, and from those who fear, not without reason, that the NRS is part of a scheme to cut Head Start or to defund existing programs and give the money instead to states and/or faith-based groups.

The Bush administration didn't need to get Congressional approval to implement the NRS in 2003, but Congress can pass legislation blocking continuation of the NRS or forcing changes in it. Differences between the parties and houses about what to do about NRS were a major stumbling block to Head Start reauthorization in the previous two Congresses. With Democrats controlling both legislative houses now, it appears likely Head Start will be reauthorized this year. Last week the House Committee on Education and Labor passed legislation to reauthorize Head Start that included a provision suspending implementation of the NRS.

It's worth noting that just because the NRS is flawed, that doesn't mean all standardized testing of young children is bad, as some opponents contend. If assessment results are used carefully and the assessment tool used is appropriate to the assessment's purpose, standardized testing can play an important role in program self-assessment, research, and even accountability for early childhood programs. For example, the preschool on whose board I serve, Appletree Early Learning Public Charter School, regularly administers to its students certain standardized assessments that help us to track their progress and also provide valuable information we can use to improve our practices and curriculum, as well as to demonstrate our effectiveness to parents, our charter authorizer, and funders. These are not pen and paper tests, which, as Alice in Eduland points out, are often foolish as a way to measure what young children know. Instead, the child sits down with a friendly adult (often someone she knows) and participates in a variety of tasks, such as selecting the correct object or picture from an array in response to a question from the assessor. And there's nothing inherently cruel about this, either. Many people I've talked to who administer these tests to children say the children often find them fun, like a game. (To reiterate: How children respond to testing depends almost entirely on the attitude of the adults administering the test. There is no natural, in-born fear of tests.)

We're going to see a lot more talk about what accountability means in early childhood education in the very near future, as states continue to expand their investments in pre-k programs. The preschool movement has focused very heavily on inputs and a regulatory approach, such as requiring teachers to have bachelor's degrees, to ensure preschool quality. But the experience of K-12 education reformers over the past several decades has shown that regulation alone doesn't get you to good outcomes for kids. Incentives matter; accountability matters. And, as a practical matter, states that invest a lot in large-scale pre-k initiatives are going to want hard evidence they're getting something to show for it. That's going to mean more conversations about how do we assess learning with young kids. It will also mean developing new tools, similar to those described here, that look neither at inputs nor outcomes but at what actually goes on in classrooms. Hopefully, it will also mean states work to create integrated longitudinal data systems that allow them to track individual children from pre-k all the way through K-12; such systems would help with the growth or value-added analysis lots of people want to see integrated into accountability models ad the K-12 level, and would also allow states to track what we REALLY want to know when we talk about whether or not preschool works: Are the kids who participate in such programs doing better 5, 10, more years down the line?